Joint estimation of biogeochemical model parameters from multiple experiments: A bayesian approach applied to mercury methylation. (September 2022)

• Record Type: Journal Article
• Authors: Rathore, Saubhagya S.; Schwartz, Grace E.; Brooks, Scott C.; Painter, Scott L.
• Is Part Of: Environmental modelling & software. Volume 155(2022)
• Publication Date: 2022-09
• Subjects: Bayesian inference -- Parameter uncertainty -- Mercury methylation; Environmental monitoring -- Computer programs -- Periodicals; Ecology -- Computer simulation -- Periodicals; Digital computer simulation -- Periodicals; Computer software -- Periodicals; Environmental Monitoring -- Periodicals; Computer Simulation -- Periodicals; Environnement -- Surveillance -- Logiciels -- Périodiques; Écologie -- Simulation, Méthodes de -- Périodiques; Simulation par ordinateur -- Périodiques; Logiciels -- Périodiques; Computer software; Digital computer simulation; Ecology -- Computer simulation; Environmental monitoring -- Computer programs; Periodicals; Electronic journals; 363.70015118
• DOI: 10.1016/j.envsoft.2022.105453
• Languages: English
• ISSNs: 1364-8152

Abstract

Abstract: To characterize complex biogeochemical systems, results from multiple experiments, where each targets a specific subprocess, are commonly combined. The resulting datasets are interpreted through the calibration of biogeochemical models for process inference and predictions. Commonly used calibration approaches of fitting datasets from individual experiments to subprocess models one at a time is prone to missing information shared between datasets and incomplete uncertainty propagation. We propose a Bayesian joint-fitting scheme addressing the above-mentioned concerns by jointly fitting all the available datasets, thus calibrating the entire biogeochemical model in one go using Markov Chain Monte Carlo (MCMC). The identification of null spaces in the parameter distributions from MCMC guided the simplification of certain subprocess models. For example, fast kinetic sorption was replaced by equilibrium sorption, and Monod demethylation was replaced by first-order demethylation. Joint fitting of datasets resulted in complete uncertainty propagation with parameter estimates informed by all available data. Highlights: Jointly fitted sorption and methylation datasets allow information sharing and uncertainty propagation across process models. Mapping full joint distributions of parameters identified null-spaces and facilitated model simplification. Sediments collected from near-bank and near-center of the stream exhibited different sorption and methylation kinetics.